Re: Re: Xen spinlock questions

[Jeremy Fitzhardinge <jeremy@goop.org>]

Jan Beulich wrote:
>>>> Jeremy Fitzhardinge <jeremy@goop.org> 06.08.08 22:21 >>>
>>>>         
>> Jan Beulich wrote:
>>     
>>>> In my tests, I only see it fall into the slow path a couple of thousand 
>>>> times per cpu for a kernbench run.
>>>>     
>>>>         
>>> Hmm, that's different then for me. Actually, I see a significant spike at
>>> modpost stage 2, when all the .ko-s get linked. The (spinlock) interrupt
>>> rate gets up to between 1,000 and 2,000 per CPU and second.
>>>   
>>>       
>> BTW, how much contention for CPUs was there?
>>     
>
> In Xen or the kernel? So far these are all Dom0-only measurements, so
> in Xen there should be no contention.
>   

OK.  The spinlock results are most interesting when there's contention 
for physical cpus between domains.

> To answer your other mail's question: It's defconfig builds I'm using.
>   
[SuSE defconfigs]  Presumably that has a lot enabled as modules.

    J